The present invention relates generally to an optical transmission fiber that has improved characteristics for minimizing non-linear effects, and specifically to an optical fiber for use in a wavelength-division-multiplexing (WDM) system that has two refractive index peaks with the maximum index of refraction difference located in an outer core region.
In optical communication systems, non-linear optical effects are known to degrade the quality of transmission along standard transmission optical fiber in certain circumstances. These non-linear effects, which include four-wave mixing (FWM), self-phase modulation (SPM), cross-phase modulation (XPM), modulation instability (MI), stimulated Brillouin scattering (SBS) and stimulated Raman scattering (SRS), particularly cause distortion in high power systems.
The strength of non-linear effects acting on pulse propagation in optical fibers is linked to the product of the non-linearity coefficient γ and the power P. The definition of the non-linearity coefficient, as given in the paper “Nonlinear pulse propagation in a monomode dielectric guide” by Y. Kodama et al., IEEE Journal of Quantum Electronics, vol. QE-23, No. 5, 1987, is the following:                     γ        =                              1                          λ              ⁢                                                           ⁢                              n                eff                                              ⁢                                                    ∫                0                ∞                            ⁢                                                n                  ⁡                                      (                    r                    )                                                  ⁢                                                      n                    2                                    ⁡                                      (                    r                    )                                                  ⁢                                                                                                F                      ⁡                                              (                        r                        )                                                                                                  4                                ⁢                r                ⁢                                  ⅆ                  r                                                                                    [                                                      ∫                    0                    ∞                                    ⁢                                                                                                                                      F                          ⁡                                                      (                            r                            )                                                                                                                      2                                        ⁢                    r                    ⁢                                          ⅆ                      r                                                                      ]                            2                                                          (        1        )            where r is the radial coordinate of the fiber, neff is the effective mode refractive index, λ is a signal wavelength, n(r) is the refractive index radial distribution, n2(r) is the non-linear index coefficient radial distribution, and F(r) is the fundamental mode radial distribution.
Applicants have identified that equation (1) takes into account the radial dependence of the non-linear index coefficient n2 which is due to the varying concentration of the fiber dopants used to raise (or to lower) the refractive index with respect to that of pure silica.
If we neglect the radial dependence of the non-linear index coefficient n2 we obtain a commonly used expression for the coefficient γ.                     γ        =                              2            ⁢                                                   ⁢            π            ⁢                                                   ⁢                          n              2                                            λ            ⁢                                                   ⁢                          A              eff                                                          (        2        )            where we have introduced the so called effective core area, or briefly, effective area,                               A          eff                =                                            2              ⁢                                                           ⁢                                                π                  ⁢                                                                           [                                                            ∫                      0                      ∞                                        ⁢                                                                                                                                                  F                            ⁡                                                          (                              r                              )                                                                                                                                2                                            ⁢                      r                      ⁢                                              ⅆ                        r                                                                              ]                                2                                                                    ∫                0                ∞                            ⁢                                                                                                              F                      ⁡                                              (                        r                        )                                                                                                  4                                ⁢                r                ⁢                                  ⅆ                  r                                                              .                                    (        3        )            The approximation (2), in contrast to the definition (1) does not distinguish between refractive index radial profiles that have the same effective core area Aeff value but different γ values. While 1/Aeff is often used as a measure of the strength of non-linear effects in a transmission fiber, γ as defined by equation (1) actually provides a better measure of the strength of those effects.
Group velocity dispersion also provides a limitation to quality transmission of optical signals across long distances. Group velocity dispersion broadens an optical pulse during its transmission across long distances, which may lead to dispersion of the optical energy outside a time slot assigned for the pulse. Although dispersion of an optical pulse can be somewhat avoided by decreasing the spacing between regenerators in a transmission system, this approach is costly and does not allow one to exploit the advantages of repeaterless optical amplification.
One known way of counteracting dispersion is by adding suitable dispersion compensating devices, such as gratings or dispersion compensating fibers, to the telecommunication system.
Furthermore, to compensate dispersion, one trend in optical communications is toward the use of soliton pulses, a particular type of RZ (Return-to-Zero) modulation signal, that maintain their pulse width over longer distances by balancing the effects of group velocity dispersion with the non-linear phenomenon of self-phase modulation. The basic relation that governs soliton propagation in a single mode optical fiber is the following:                                           P            0                    ⁢                      T            0            2                          =                  cos          ⁢                                           ⁢          t          ⁢                                           ⁢                                    D              ⁢                                                           ⁢                              λ                2                                      γ                                              (        4        )            where P0 is the peak power of a soliton pulse, T0 is the time duration of the pulse, D is the total dispersion, λ is the center wavelength of the soliton signal, and γ is the previously introduced fiber non-linearity coefficient. Satisfaction of equation (4) is necessary in order for a pulse to be maintained in a soliton condition during propagation.
A possible problem that arises in the transmission of solitons in accordance with equation (4) is that a conventional optical transmission fiber is lossy, which causes the peak power P0 of the soliton pulse to decrease exponentially along the length of the fiber between optical amplifiers. To compensate for this decrease, one can set the soliton power P0 at its launch point at a value sufficient to compensate for the subsequent decrease in power along the transmission line. An alternative approach, as disclosed for example in F. M. Knox et al., paper WeC.3.2, page 3.101-104, ECOC '96, Oslo (Norway), is to compensate (with dispersion compensating fiber, although fibre Bragg gratings can also be used) for the dispersion accumulated by the pulses along the stretches of the transmission line where the pulses' peak power is below a soliton propagation condition.
Optical fibers having a low non-linearity coefficient are preferred for use in transmission systems, such as Non-Return-to-Zero (NRZ) optically amplified WDM systems, as well as non amplified systems, to avoid or limit the non-linear effects mentioned above. Furthermore, fibers with a lower non-linearity coefficient allow an increase in the launch power while maintaining non-linear effects at the same level. An increased launch power in turn means a better SIN ratio at the receiver (lower BER) and/or the possibility to reach longer transmission distances by increasing the amplifier spacing. Accordingly, Applicants have addressed a need for optical fibers having low values of non-linearity coefficient γ.
Also in the case of soliton systems, to increase the spacing between amplifiers one can increase the launch power for the pulses using more powerful amplifiers. In this case, however, equation (4) implies that if the launch power is increased and the soliton pulse duration remains constant, the ratio Dλ2/γ must accordingly be increased. Therefore, lower values of non-linear coefficient γ are desirable also to provide an increased distance between line amplifiers in a soliton transmission system.
Patents and publications have discussed the design of optical transmission fibers using a segmented core or double-cladding refractive index profile and fibers having a large effective area. For example, U.S. Pat. No. 5,579,428 discloses a single-mode optical fiber designed for use in a WDM soliton telecommunication system using optical lumped or distributed amplifiers. Over a preselected wavelength range, the total dispersion for the disclosed optical fiber lies within a preselected range of positive values high enough to balance self-phase modulation for WDM soliton propagation. As well, the dispersion slope lies within a preselected range of values low enough to prevent collisions between WDM solitons and to reduce their temporal and spectral shifts. The proposed fiber of the '428 patent is a segmented core with a region of maximum index of refraction in the core of the fiber.
U.S. Pat. No. 4,715,679 discloses an optical fiber having a segmented core of a depressed refractive index for making low dispersion, low loss waveguides. The '679 patent discloses a plurality of refractive index profiles including an idealized profile having an area of maximum index of refraction at an annular region outside the inner core of the fiber but inside an outer core annular region.
U.S. Pat. No. 4,877,304 discloses an optical fiber that has a core profile with a maximum refractive index greater than that of its cladding. U.S. Pat. No. 4,889,404 discloses an asymmetrical bi-directional optical communication system including an optical fiber. While the '304 and '404 patents also describe idealized refractive index profiles potentially having an outer annular region with an increased index of refraction, no specific examples corresponding to those profiles are disclosed and the patents are silent as to the non-linear characteristics of optical fibers having those profiles.
U.S. Pat. No. 5,684,909, EP 789,255, and EP 724,171 disclose single mode optical fibers having large effective areas made by a segmented refractive index core profile. This patent and applications describe computer simulations for obtaining fibers with a large effective area for use in long distance, high bit rate optical systems. The '909 patent shows a core profile having two non-adjacent profile segments having a positive index of refraction and two additional non-adjacent segments having a negative index of refraction. The '909 patent aims to achieve a fiber with a substantially zero dispersion slope from the segmented core profile. The fibers disclosed in EP 789,255 have extremely large effective areas achieved by a refractive index profile with a segmented core but having at least two non-adjacent segments with negative refractive difference. EP 724,171 discloses optical fibers with the maximum index of refraction present at the center of the fiber.
U.S. Pat. No. 5,555,340 discloses a dispersion compensating optical fiber having a segmented core for obtaining dispersion compensation. The '340 patent discloses a refractive index profile where a resin film surrounding a cladding has a higher index of refraction than the inner core of the fiber. This resin, however, does not serve as a low-loss light-conductive layer in the fiber structure.